Uber is trying to own the part of robotaxis nobody wants to own

Everyone is arguing about whether Waymo, Tesla or another company will build the best robotaxi.

Uber may be playing a different game.

It doesn’t need to build the best autonomous car.

It may be trying to become the company that handles everything around the car.

The distinction matters.

A robotaxi isn’t just a vehicle with an AI driver.

Someone has to finance it.

Someone has to insure it.

Someone has to charge it.

Someone has to maintain it.

Someone has to manage the fleet.

Someone has to deal with regulators.

Someone has to match vehicles with riders.

Someone has to handle payments, customer service and the ugly operational problems that appear when hundreds or thousands of vehicles are moving around a city.

Uber is quietly moving into almost all of those layers.

The company says it expects to commit more than $10 billion across AV investments, infrastructure and vehicle offtake commitments. Its partners have committed roughly 120,000 vehicles to the Uber network over the coming years.

But Uber isn’t necessarily buying 120,000 cars and putting them on its own balance sheet.

That’s the interesting part.

Pony.ai’s European expansion provides a clean example.

Pony supplies the autonomous-driving technology.

A local fleet company owns and operates the vehicles.

Uber supplies the marketplace, booking, payment and customer relationship.

Different companies own different pieces.

Uber sits in the middle.

Uber has also created a dedicated autonomous-vehicle insurance facility with Marsh and Apollo’s ibott to provide liability coverage for AV partners.

That sounds boring until you realize insurance is one of the things that can determine whether a robotaxi fleet can actually operate commercially.

Then there is financing.

Uber says it is exploring outside capital providers to finance future AV deployments while preserving its own balance sheet.

That’s a very different model from “Uber is buying robotaxis.”

It is closer to building the financial and operating plumbing that lets other companies put robotaxis on the road.

And the company’s existing business gives it a huge testing ground.

Uber had 208 million monthly active platform consumers in the second quarter.

It handled 3.9 billion trips in three months.

Gross bookings reached $58 billion.

Free cash flow exceeded $10 billion over the previous twelve months for the first time.

That means Uber can experiment with AVs without having to bet the company on a single autonomous-driving technology.

Waymo can win.

Pony can win.

WeRide can win.

Wayve can win.

Uber can potentially work with all of them.

That’s also why the comment war over whether riders will download another app misses something.

Yes, they will.

There is nothing sacred about the Uber icon on a phone.

If another company offers a better ride for less money, people can download another app in seconds.

That is the obvious weakness in the “distribution moat” argument.

Uber knows this.

Its own strategy increasingly looks less like protecting an app and more like becoming an operating layer for a fragmented robotaxi industry.

That’s a much harder thing to replace.

Replacing an app is easy.

Replacing the insurance arrangement, fleet operation, financing relationships, regulatory infrastructure, payment system, dispatch technology and millions of existing riders is a different project.

And Uber is already building the hybrid version.

Human drivers and autonomous vehicles operate on the same marketplace.

In August, Uber launched autonomous rides in Zagreb through Pony.ai and Verne. Pony provides the technology. Verne owns and operates the fleet. Uber handles the customer-facing marketplace. Uber says human drivers and AVs are expected to coexist for the foreseeable future.

That gives Uber an unusual hedge.

If robotaxis don’t become economical, Uber still has its human-driver business.

If robotaxis do become economical, Uber can add them without replacing the entire existing network overnight.

If dozens of AV manufacturers compete, Uber can potentially host multiple winners.

And if fleet ownership becomes a terrible business because vehicles are expensive, heavily used assets with insurance and maintenance costs, Uber can leave much of that headache with the fleet operators.

That’s the part of the story I would pay attention to.

The robotaxi companies are fighting over the vehicle.

Uber is trying to build the machinery that makes the vehicle commercially useful.

And that could create a strange outcome.

The company with the best autonomous driver may not control the economics of the robotaxi market.

The company that controls the boring stuff around the autonomous driver might.

Not financial advice.